Society of Critical Care Medicine and the
Infectious Diseases Society of America
Guidelines for Evaluating New Fever in
Downloaded from http://journals.lww.com/ccmjournal by BhDMf5ePHKav1zEoum1tQfN4a+kJLhEZgbsIHo4XMi0hCy
Adult Patients in the ICU
wCX1AWnYQp/IlQrHD3i3D0OdRyi7TvSFl4Cf3VC4/OAVpDDa8K2+Ya6H515kE= on 10/12/2023
Naomi P. O’Grady, MD, FCCM,
RATIONALE: Fever is frequently an early indicator of infection and often requires FIDSA1
rigorous diagnostic evaluation. Earnest Alexander, PharmD,
OBJECTIVES: This is an update of the 2008 Infectious Diseases Society of FCCM2
America and Society (IDSA) and Society of Critical Care Medicine (SCCM) Waleed Alhazzani, MBBS, MSc,
guideline for the evaluation of new-onset fever in adult ICU patients without severe FRCPC3
immunocompromise, now using the Grading of Recommendations Assessment, Fayez Alshamsi, MBBS4
Development, and Evaluation (GRADE) methodology. Jennifer Cuellar-Rodriguez, MD5
PANEL DESIGN: The SCCM and IDSA convened a taskforce to update the 2008 Brian K. Jefferson, DNP,
version of the guideline for the evaluation of new fever in critically ill adult patients, ACNP-BC, FCCM6
which included expert clinicians as well as methodologists from the Guidelines in
Andre C. Kalil, MD, MPH, FCCM,
Intensive Care, Development and Evaluation Group. The guidelines committee con- FIDSA7
sisted of 12 experts in critical care, infectious diseases, clinical microbiology, organ
Stephen M. Pastores, MD, MACP,
transplantation, public health, clinical research, and health policy and administration. All FCCP, FCCM8
task force members followed all conflict-of-interest procedures as documented in the
Robin Patel, MD, FIDSA,
American College of Critical Care Medicine/SCCM Standard Operating Procedures
FRCPC9,10
Manual and the IDSA. There was no industry input or funding to produce this guideline.
David van Duin, MD, PhD,
METHODS: We conducted a systematic review for each population, interven- FIDSA11
tion, comparison, and outcomes question to identify the best available evidence,
David J. Weber, MD, FIDSA,
statistically summarized the evidence, and then assessed the quality of evidence FSHEA, FRSM, FAST11
using the GRADE approach. We used the evidence-to-decision framework to for-
Stanley Deresinski, MD, FIDSA12
mulate recommendations as strong or weak or as best-practice statements.
RESULTS: The panel issued 12 recommendations and 9 best practice statements.
The panel recommended using central temperature monitoring methods, including
thermistors for pulmonary artery catheters, bladder catheters, or esophageal bal-
loon thermistors when these devices are in place or accurate temperature measure-
ments are critical for diagnosis and management. For patients without these devices
in place, oral or rectal temperatures over other temperature measurement methods
that are less reliable such as axillary or tympanic membrane temperatures, nonin-
vasive temporal artery thermometers, or chemical dot thermometers were recom-
mended. Imaging studies including ultrasonography were recommended in addition
to microbiological evaluation using rapid diagnostic testing strategies. Biomarkers
were recommended to assist in guiding the discontinuation of antimicrobial therapy.
All recommendations issued were weak based on the quality of data.
CONCLUSIONS: The guidelines panel was able to formulate several recommenda-
tions for the evaluation of new fever in a critically ill adult patient, acknowledging that
most recommendations were based on weak evidence. This highlights the need for
the rapid advancement of research in all aspects of this issue—including better non- Copyright © 2023 by the Society of
invasive methods to measure core body temperature, the use of diagnostic imaging, Critical Care Medicine and Wolters
advances in microbiology including molecular testing, and the use of biomarkers. Kluwer Health, Inc. All Rights
KEY WORDS: diagnosis; evaluation; fever; guidelines; infection; temperature Reserved.
measurement DOI: 10.1097/CCM.0000000000006022
1570 www.ccmjournal.org November 2023 • Volume 51 • Number 11
Copyright © 2023 by the Society of Critical Care Medicine and Wolters Kluwer Health, Inc. All Rights Reserved.
, Special Articles
Fever, a frequent early indicator of infection, occurs in This is an update, now using the Grading of
26–88% of adult ICU patients, depending on the def- Recommendations Assessment, Development, and
inition used and characteristics of the cohort studied Evaluation (GRADE) methodology, of the 2008
(1). The range of potential etiologies of fever is vast and Infectious Diseases Society of America and Society
includes both infectious and noninfectious causes (2). (IDSA) and Society of Critical Care Medicine
Noninfectious causes of fever should be considered in (SCCM) guideline for the evaluation of new-onset
Downloaded from http://journals.lww.com/ccmjournal by BhDMf5ePHKav1zEoum1tQfN4a+kJLhEZgbsIHo4XMi0hCy
the differential diagnosis (Table 1), but because early fever in adult ICU patients without severe immu-
treatment initiation may improve outcomes of infec- nocompromise (3). Any recommendation from the
wCX1AWnYQp/IlQrHD3i3D0OdRyi7TvSFl4Cf3VC4/OAVpDDa8K2+Ya6H515kE= on 10/12/2023
tions, initial evaluation of patients with new-onset 2008 guideline not specifically addressed in this up-
fever is usually directed at potential microbial causes, date remains in place. In this document, we address
and this is the primary focus of this guideline. microbiologic studies, imaging procedures, and
the use of biomarkers in the diagnostic evaluation
of fever with initial onset after ICU admission, fo-
TABLE 1. cusing on detection of potential infectious etiolo-
Noninfectious Causes of New Fever in ICU gies. It should be noted that not all febrile episodes
Patients dictate a need for investigation, that is, those in
which a noninfectious etiology is obvious such as
Acalculous Cholecystitis fever occurring immediately in the postoperative
Acute myocardial infarction state. For those fevers that do require investigation, a
Adrenal insufficiency good history and physical examination will often re-
Atelectasis veal potential sources of infection. Diagnostic stud-
Blood product transfusion
ies should then be sent with those potential sources
in focus rather than reflexively sending cultures for
Cytokine release syndrome
all possible sources. Although much of this docu-
Dressler syndrome (pericardial injury syndrome)
ment and its recommendations may be applicable
Drug fever
to severely immunocompromised patients, such
Fat emboli as organ transplant recipients and those with se-
Fibroproliferative phase of acute respiratory distress syndrome vere neutropenia, these populations are not directly
Gout addressed here. The variability and complexities of
Heterotopic ossification different types of immunocompromise make this a
Immune reconstitution inflammatory syndrome task that cannot be accomplished in the context of a
Intracranial bleed generally applicable guideline.
Jarisch-Herxheimer reaction The guideline is intended for use by members of
Malignant hyperthermia multidisciplinary care teams managing mixed popu-
Neuroleptic malignant syndrome lations of critically ill patients in the ICU, including
Nonconvulsive status epilepticus
intensivists, infectious diseases specialists, advanced
practice providers, clinical pharmacists, nurses, respi-
Pancreatitis
ratory therapists, and policymakers.
Pulmonary infarction
Pneumonitis without infection METHODOLOGY
Serotonin syndrome
Stroke
The SCCM and IDSA reconvened a taskforce to update
the 2008 version of the guideline for evaluation of new
Thyroid storm
fever in critically ill adult patients (3). The taskforce
Transplant rejection
included expert clinicians as well as methodologists
Tumor lysis syndrome
from the Guidelines in Intensive Care, Development
Venous thrombosis and Evaluation Group. The guidelines committee
Withdrawal from certain substances including alcohol, consisted of 12 experts in critical care, infectious dis-
opiates, barbiturates, benzodiazepines
eases, clinical microbiology, organ transplantation,
Critical Care Medicine www.ccmjournal.org 1571
Copyright © 2023 by the Society of Critical Care Medicine and Wolters Kluwer Health, Inc. All Rights Reserved.
, O’Grady et al
public health, clinical research, and health policy and University; http://www.cebm.brown.edu/openmeta/
administration. doc/openMA_help.html#self) (4, 5).
The panel was divided into five subgroups focusing Risk of bias for individual RCTs was assessed using
on fever determination, treatment, imaging, micro- the Cochrane Collaboration risk of bias assessment
biological evaluation, and rapid diagnostic testing. tool (6).
Group leaders were responsible to develop Population, Guideline methodologists assessed the quality
Downloaded from http://journals.lww.com/ccmjournal by BhDMf5ePHKav1zEoum1tQfN4a+kJLhEZgbsIHo4XMi0hCy
Intervention, Comparison, and Outcomes (PICOs) of the body of evidence using the GRADE meth-
questions for their group. The final list of PICO ques- odology and rated quality as high, moderate, low,
wCX1AWnYQp/IlQrHD3i3D0OdRyi7TvSFl4Cf3VC4/OAVpDDa8K2+Ya6H515kE= on 10/12/2023
tions was approved by consensus of the taskforce or very low based on the following domains: risk of
members. The taskforce members then provided all po- bias, inconsistency, indirectness, imprecision, publi-
tentially relevant outcomes for each PICO. Electronic cation bias, and other criteria (7, 8). Methodologists
voting was used to prioritize outcomes according to used GRADEpro guideline development tool online
importance to patients and clinicians on a scale from software (https://gradepro.org/) to produce evidence
1 to 9 (not important to critical); only outcomes with summary tables (9).
an average score of 7 and above were selected. In all, 26 We used the evidence-to-decision (EtD) framework
PICOs were included (supplement, http://links.lww. to formulate recommendations. Recommendations
com/CCM/H386). and their respective EtD are provided in supple-
All task force members followed all conflict-of- ment (http://links.lww.com/CCM/H386). The task
interest procedures as documented in the American force met monthly and completed the electronic EtD
College of Critical Care Medicine (ACCM)/SCCM forms for each PICO to formulate a recommenda-
Standard Operating Procedures Manual and the IDSA. tion. Factors considered when determining the direc-
There was no industry input or funding to produce this tion and strength of recommendation were: quality
guideline. of evidence, magnitude of effect, patient values and
A professional medical librarian developed the preferences, resources, cost, acceptability, and feasi-
search strategies for the PICO questions. We searched bility (10). For strong recommendations, we used the
Cochrane Central and MEDLINE databases for rele- wording “we recommend” and for weak recommen-
vant studies published in the English language from dations, we used “we suggest.” When the taskforce
inception through December 2018. We updated the was confident that desirable outweighed undesirable
searches through June 2022 just before our final sub- effects, a strong recommendation for an intervention
mission as recommended by the ACCM. was made while a strong recommendation against an
We aimed to include recent (10 yr old or less) rel- intervention was made when undesirable outweighed
evant systematic reviews or update outdated reviews desirable effects. A weak recommendation for or
when newer data were available. In the absence of a against an intervention was made when the task-
published meta-analysis for a specific PICO question, force was less confident about the balance between
we used the random-effects model to pool the effect desirable and undesirable effects, respectively. Best-
sizes across randomized controlled trials (RCTs) or practice statements were made only when suggested
observational studies, when applicable. For inter- GRADE criteria for best-practice statements were
ventions, we presented the pooled results as relative met (11). Best practice statements had to be clear,
risk (RR) and 95% CI for binary outcomes and mean answer an important actionable question where the
difference (MD) and 95% CI for continuous out- benefit would be unequivocally large, and evidence
comes. All analyses were performed using RevMan would be difficult to collect or summarize. An explicit
software (Review Manager [RevMan] Version 5.4, rationale for the benefit was provided.
The Cochrane Collaboration, 2020). For diagnostic After finalizing a preliminary recommendation,
tests, we used the random-effects bivariate anal- members of the taskforce received electronic links to
ysis and hierarchical standard receiver operating indicate their agreement or disagreement. Consensus
curve (ROC) methods to present pooled sensitivity required 80% agreement by at least 75% of the voting
and specificity using Open Meta-Analyst software panel. Recommendations and best-practice statements
(Center for Evidence Synthesis in Health, Brown are listed in Table 2.
1572 www.ccmjournal.org November 2023 • Volume 51 • Number 11
Copyright © 2023 by the Society of Critical Care Medicine and Wolters Kluwer Health, Inc. All Rights Reserved.
Infectious Diseases Society of America
Guidelines for Evaluating New Fever in
Downloaded from http://journals.lww.com/ccmjournal by BhDMf5ePHKav1zEoum1tQfN4a+kJLhEZgbsIHo4XMi0hCy
Adult Patients in the ICU
wCX1AWnYQp/IlQrHD3i3D0OdRyi7TvSFl4Cf3VC4/OAVpDDa8K2+Ya6H515kE= on 10/12/2023
Naomi P. O’Grady, MD, FCCM,
RATIONALE: Fever is frequently an early indicator of infection and often requires FIDSA1
rigorous diagnostic evaluation. Earnest Alexander, PharmD,
OBJECTIVES: This is an update of the 2008 Infectious Diseases Society of FCCM2
America and Society (IDSA) and Society of Critical Care Medicine (SCCM) Waleed Alhazzani, MBBS, MSc,
guideline for the evaluation of new-onset fever in adult ICU patients without severe FRCPC3
immunocompromise, now using the Grading of Recommendations Assessment, Fayez Alshamsi, MBBS4
Development, and Evaluation (GRADE) methodology. Jennifer Cuellar-Rodriguez, MD5
PANEL DESIGN: The SCCM and IDSA convened a taskforce to update the 2008 Brian K. Jefferson, DNP,
version of the guideline for the evaluation of new fever in critically ill adult patients, ACNP-BC, FCCM6
which included expert clinicians as well as methodologists from the Guidelines in
Andre C. Kalil, MD, MPH, FCCM,
Intensive Care, Development and Evaluation Group. The guidelines committee con- FIDSA7
sisted of 12 experts in critical care, infectious diseases, clinical microbiology, organ
Stephen M. Pastores, MD, MACP,
transplantation, public health, clinical research, and health policy and administration. All FCCP, FCCM8
task force members followed all conflict-of-interest procedures as documented in the
Robin Patel, MD, FIDSA,
American College of Critical Care Medicine/SCCM Standard Operating Procedures
FRCPC9,10
Manual and the IDSA. There was no industry input or funding to produce this guideline.
David van Duin, MD, PhD,
METHODS: We conducted a systematic review for each population, interven- FIDSA11
tion, comparison, and outcomes question to identify the best available evidence,
David J. Weber, MD, FIDSA,
statistically summarized the evidence, and then assessed the quality of evidence FSHEA, FRSM, FAST11
using the GRADE approach. We used the evidence-to-decision framework to for-
Stanley Deresinski, MD, FIDSA12
mulate recommendations as strong or weak or as best-practice statements.
RESULTS: The panel issued 12 recommendations and 9 best practice statements.
The panel recommended using central temperature monitoring methods, including
thermistors for pulmonary artery catheters, bladder catheters, or esophageal bal-
loon thermistors when these devices are in place or accurate temperature measure-
ments are critical for diagnosis and management. For patients without these devices
in place, oral or rectal temperatures over other temperature measurement methods
that are less reliable such as axillary or tympanic membrane temperatures, nonin-
vasive temporal artery thermometers, or chemical dot thermometers were recom-
mended. Imaging studies including ultrasonography were recommended in addition
to microbiological evaluation using rapid diagnostic testing strategies. Biomarkers
were recommended to assist in guiding the discontinuation of antimicrobial therapy.
All recommendations issued were weak based on the quality of data.
CONCLUSIONS: The guidelines panel was able to formulate several recommenda-
tions for the evaluation of new fever in a critically ill adult patient, acknowledging that
most recommendations were based on weak evidence. This highlights the need for
the rapid advancement of research in all aspects of this issue—including better non- Copyright © 2023 by the Society of
invasive methods to measure core body temperature, the use of diagnostic imaging, Critical Care Medicine and Wolters
advances in microbiology including molecular testing, and the use of biomarkers. Kluwer Health, Inc. All Rights
KEY WORDS: diagnosis; evaluation; fever; guidelines; infection; temperature Reserved.
measurement DOI: 10.1097/CCM.0000000000006022
1570 www.ccmjournal.org November 2023 • Volume 51 • Number 11
Copyright © 2023 by the Society of Critical Care Medicine and Wolters Kluwer Health, Inc. All Rights Reserved.
, Special Articles
Fever, a frequent early indicator of infection, occurs in This is an update, now using the Grading of
26–88% of adult ICU patients, depending on the def- Recommendations Assessment, Development, and
inition used and characteristics of the cohort studied Evaluation (GRADE) methodology, of the 2008
(1). The range of potential etiologies of fever is vast and Infectious Diseases Society of America and Society
includes both infectious and noninfectious causes (2). (IDSA) and Society of Critical Care Medicine
Noninfectious causes of fever should be considered in (SCCM) guideline for the evaluation of new-onset
Downloaded from http://journals.lww.com/ccmjournal by BhDMf5ePHKav1zEoum1tQfN4a+kJLhEZgbsIHo4XMi0hCy
the differential diagnosis (Table 1), but because early fever in adult ICU patients without severe immu-
treatment initiation may improve outcomes of infec- nocompromise (3). Any recommendation from the
wCX1AWnYQp/IlQrHD3i3D0OdRyi7TvSFl4Cf3VC4/OAVpDDa8K2+Ya6H515kE= on 10/12/2023
tions, initial evaluation of patients with new-onset 2008 guideline not specifically addressed in this up-
fever is usually directed at potential microbial causes, date remains in place. In this document, we address
and this is the primary focus of this guideline. microbiologic studies, imaging procedures, and
the use of biomarkers in the diagnostic evaluation
of fever with initial onset after ICU admission, fo-
TABLE 1. cusing on detection of potential infectious etiolo-
Noninfectious Causes of New Fever in ICU gies. It should be noted that not all febrile episodes
Patients dictate a need for investigation, that is, those in
which a noninfectious etiology is obvious such as
Acalculous Cholecystitis fever occurring immediately in the postoperative
Acute myocardial infarction state. For those fevers that do require investigation, a
Adrenal insufficiency good history and physical examination will often re-
Atelectasis veal potential sources of infection. Diagnostic stud-
Blood product transfusion
ies should then be sent with those potential sources
in focus rather than reflexively sending cultures for
Cytokine release syndrome
all possible sources. Although much of this docu-
Dressler syndrome (pericardial injury syndrome)
ment and its recommendations may be applicable
Drug fever
to severely immunocompromised patients, such
Fat emboli as organ transplant recipients and those with se-
Fibroproliferative phase of acute respiratory distress syndrome vere neutropenia, these populations are not directly
Gout addressed here. The variability and complexities of
Heterotopic ossification different types of immunocompromise make this a
Immune reconstitution inflammatory syndrome task that cannot be accomplished in the context of a
Intracranial bleed generally applicable guideline.
Jarisch-Herxheimer reaction The guideline is intended for use by members of
Malignant hyperthermia multidisciplinary care teams managing mixed popu-
Neuroleptic malignant syndrome lations of critically ill patients in the ICU, including
Nonconvulsive status epilepticus
intensivists, infectious diseases specialists, advanced
practice providers, clinical pharmacists, nurses, respi-
Pancreatitis
ratory therapists, and policymakers.
Pulmonary infarction
Pneumonitis without infection METHODOLOGY
Serotonin syndrome
Stroke
The SCCM and IDSA reconvened a taskforce to update
the 2008 version of the guideline for evaluation of new
Thyroid storm
fever in critically ill adult patients (3). The taskforce
Transplant rejection
included expert clinicians as well as methodologists
Tumor lysis syndrome
from the Guidelines in Intensive Care, Development
Venous thrombosis and Evaluation Group. The guidelines committee
Withdrawal from certain substances including alcohol, consisted of 12 experts in critical care, infectious dis-
opiates, barbiturates, benzodiazepines
eases, clinical microbiology, organ transplantation,
Critical Care Medicine www.ccmjournal.org 1571
Copyright © 2023 by the Society of Critical Care Medicine and Wolters Kluwer Health, Inc. All Rights Reserved.
, O’Grady et al
public health, clinical research, and health policy and University; http://www.cebm.brown.edu/openmeta/
administration. doc/openMA_help.html#self) (4, 5).
The panel was divided into five subgroups focusing Risk of bias for individual RCTs was assessed using
on fever determination, treatment, imaging, micro- the Cochrane Collaboration risk of bias assessment
biological evaluation, and rapid diagnostic testing. tool (6).
Group leaders were responsible to develop Population, Guideline methodologists assessed the quality
Downloaded from http://journals.lww.com/ccmjournal by BhDMf5ePHKav1zEoum1tQfN4a+kJLhEZgbsIHo4XMi0hCy
Intervention, Comparison, and Outcomes (PICOs) of the body of evidence using the GRADE meth-
questions for their group. The final list of PICO ques- odology and rated quality as high, moderate, low,
wCX1AWnYQp/IlQrHD3i3D0OdRyi7TvSFl4Cf3VC4/OAVpDDa8K2+Ya6H515kE= on 10/12/2023
tions was approved by consensus of the taskforce or very low based on the following domains: risk of
members. The taskforce members then provided all po- bias, inconsistency, indirectness, imprecision, publi-
tentially relevant outcomes for each PICO. Electronic cation bias, and other criteria (7, 8). Methodologists
voting was used to prioritize outcomes according to used GRADEpro guideline development tool online
importance to patients and clinicians on a scale from software (https://gradepro.org/) to produce evidence
1 to 9 (not important to critical); only outcomes with summary tables (9).
an average score of 7 and above were selected. In all, 26 We used the evidence-to-decision (EtD) framework
PICOs were included (supplement, http://links.lww. to formulate recommendations. Recommendations
com/CCM/H386). and their respective EtD are provided in supple-
All task force members followed all conflict-of- ment (http://links.lww.com/CCM/H386). The task
interest procedures as documented in the American force met monthly and completed the electronic EtD
College of Critical Care Medicine (ACCM)/SCCM forms for each PICO to formulate a recommenda-
Standard Operating Procedures Manual and the IDSA. tion. Factors considered when determining the direc-
There was no industry input or funding to produce this tion and strength of recommendation were: quality
guideline. of evidence, magnitude of effect, patient values and
A professional medical librarian developed the preferences, resources, cost, acceptability, and feasi-
search strategies for the PICO questions. We searched bility (10). For strong recommendations, we used the
Cochrane Central and MEDLINE databases for rele- wording “we recommend” and for weak recommen-
vant studies published in the English language from dations, we used “we suggest.” When the taskforce
inception through December 2018. We updated the was confident that desirable outweighed undesirable
searches through June 2022 just before our final sub- effects, a strong recommendation for an intervention
mission as recommended by the ACCM. was made while a strong recommendation against an
We aimed to include recent (10 yr old or less) rel- intervention was made when undesirable outweighed
evant systematic reviews or update outdated reviews desirable effects. A weak recommendation for or
when newer data were available. In the absence of a against an intervention was made when the task-
published meta-analysis for a specific PICO question, force was less confident about the balance between
we used the random-effects model to pool the effect desirable and undesirable effects, respectively. Best-
sizes across randomized controlled trials (RCTs) or practice statements were made only when suggested
observational studies, when applicable. For inter- GRADE criteria for best-practice statements were
ventions, we presented the pooled results as relative met (11). Best practice statements had to be clear,
risk (RR) and 95% CI for binary outcomes and mean answer an important actionable question where the
difference (MD) and 95% CI for continuous out- benefit would be unequivocally large, and evidence
comes. All analyses were performed using RevMan would be difficult to collect or summarize. An explicit
software (Review Manager [RevMan] Version 5.4, rationale for the benefit was provided.
The Cochrane Collaboration, 2020). For diagnostic After finalizing a preliminary recommendation,
tests, we used the random-effects bivariate anal- members of the taskforce received electronic links to
ysis and hierarchical standard receiver operating indicate their agreement or disagreement. Consensus
curve (ROC) methods to present pooled sensitivity required 80% agreement by at least 75% of the voting
and specificity using Open Meta-Analyst software panel. Recommendations and best-practice statements
(Center for Evidence Synthesis in Health, Brown are listed in Table 2.
1572 www.ccmjournal.org November 2023 • Volume 51 • Number 11
Copyright © 2023 by the Society of Critical Care Medicine and Wolters Kluwer Health, Inc. All Rights Reserved.
